* M9: Boards: mind maps in the document and engine (#2226)
Part of #2226, step 1 of 4. A mind map node is a board object of its own (BoardItem::MindNode: a frame hugging its text, no box) and a branch is a connector marked `branch`; the tree is read from the branches alone. `board.mindMap` places a first node with three children and `board.mindMapAdd` adds a child or a sibling, each one undo step. The normalize pass (`board::settle_mind_maps`, `figcraft_doc::mind_layout`) lays a touched map out again from its first node: 88 between columns, 48 between siblings, each stack centred on its parent; a branch that no longer joins two layers of its page is removed. Branches are curved, stroke 4, #757575, round caps. Numbers are from a FigJam map saved and decoded. No toolbar entry, shortcuts or import yet.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.1 designed; Sonnet 5.5 wrote at medium
* M12: Mind maps: a map around its own container stays put; a moved branch re-flows its old map (#2226)
A map holding a layer and something inside it is skipped by the settle pass (it never settled). Every touched branch now stirs its old ends, so re-attaching a branch's start re-flows the map it left. Boards without branches return early, a map over MIND_MAX_NODES is left whole, shift skips non-finite moves, and the doc tests share one six-node helper.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.1 reviewed; Sonnet 5.5 wrote at medium
* M12: Mind maps: FigCraft's own starter texts (#2226)
A new map's starter children read "Idea 1", "Idea 2" and "Idea 3", as the design specified,
instead of the incumbent's wording, which the first commit had copied from the measurement.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove and wrote
* M12: Docs: the mind-map revision rows go in the revision history (#2226)
The 2026-11-10 rows sat in ROADMAP.md's Headline numbers table and docs/roadmap.md's Alpha gate
table; each now opens its file's revision history.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove and wrote
---------
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
release.yml also runs nightly on main with craft-common-actions' dry-run: every release build
except macOS, with nothing signed, pushed or drafted. Breakage that only the release build hits
(riscv64, Flatpak, FreeBSD, packaging) then shows up the next morning instead of on release
day, as it did on 11 Nov 2026.
insert_model cleared the copied start with one set_para_lists call per newline, each scanning the whole text, and set_para_lists also zeroed the level of plain paragraphs. Add TextData::clear_list_starts (one pass, touches only start_offset, no-op when no paragraph has a start) and call it once over the created paragraphs, only when the text has a newline.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Opus 5.5 briefed; Sonnet 5.5 wrote at medium
* M5: Locked layers follow Figma's rules: no nudge, a locked parent's children stay locked (#1560)
Pins Figma's locked-layer behaviour (measured in Figma's web app, 2026-11-10; help center "Lock and unlock layers") in a new lock_behaviour_tests suite. Fixes the divergences: transform.nudge and transform.nudgeSize skip locked layers, object.toggleLock without a value counts a child of a locked parent as locked (so it never locks the child on its own), and the Layers panel padlock on such a child is inert.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.1 briefed; Sonnet 5.5 wrote at medium
* Layers panel: a padlock sweep skips the children of a locked layer (#1560)
A padlock drag sweep no longer collects rows with `locked_above`, so
dragging across a locked frame's children leaves their own flags alone
and draws no swept value on them. An eye sweep is unchanged.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Opus 5.5 briefed; Sonnet 5.5 wrote at medium
* M5: Lock/Unlock leaves a locked layer's child alone in a mixed selection (#1560)
With a locked frame's child and an unlocked layer selected, Lock/Unlock set both own flags, so
the child stayed locked once the frame was unlocked. With no explicit `locked` value, the command
now skips a layer whose parent is locked, as its padlock does; an explicit value (scripts) still
sets it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove and wrote
---------
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <FredApps@users.noreply.github.com>
* fix(flatpak): allow release checks and self-update helpers
* test: keep Flatpak permission parsing within strict Clippy rules
* chore: bump version to 0.17.1 for the Flatpak update fix
* pathops: check closed inputs before faces normalises, share the dash allowance across subpaths, index point incidence once
* pathops: leave out move-only subpaths when closing, budget dashes by the fitted pattern, index only join points
Enter copied the item's start (a typed "5. " gives 4) onto the new item, so once a middle item was unlisted the next one restarted at its copied start. Every paragraph an insertion creates now has start_offset 0; the first item keeps its start. Measured in Figma: unlisting "b" in 5. 6. 7. leaves "c" at 1., and the same for a list started with 1.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Opus 5.5 briefed; Sonnet 5.5 wrote at medium
The desktop app built eframe without its `links` feature, so egui-winit dropped every
`ctx.open_url` with only a log warning: the update prompt's What's new link and Download
button, the About and Credits links, Terms in Live and the link in the Dev Mode panel did
nothing on click. The feature adds no crate: `webbrowser` is already a dependency of the app.
DeckCraft's desktop app builds eframe with it too.
`links_in_the_interface_open_in_the_browser` fails without the feature.
Co-authored-by: Oleksii Sukholita <oleksii.sukholita@gmail.com>
Adds step-security/harden-runner (audit, pinned to v2.22.4 like
craft-common-actions) as the first step of each job that did not have it,
so every workflow in the org's public repos reports its egress to
StepSecurity. Audit, not block: block-mode allow lists come from each
job's observed traffic.
Typing `5. ` at the start of a plain paragraph starts a numbered list whose first item reads 5 (Figma has no start-number field); the next item goes on with 6. Digits only, 1 to 9999; `0. `, `12a. ` and `2)` stay text (not measured), `1)` works as before. The measured restart rows (plain paragraph, bullet item) are no longer marked silent.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Opus 5.5 briefed; Sonnet 5.5 wrote at medium
The paste handler read the system clipboard synchronously (2 s timeout) even
though the paste event already carries that text, and discarded the read's
result. On the desktop, the clipboard bridge delivers `event_text` that it read
while the window thread was free; re-reading from the paste handler blocks that
same window thread - and when the app itself owns the clipboard (a self-paste),
its own data source can only be served by that very thread, so every self-paste
froze the window for the full 2 s read timeout on Linux/Wayland.
Prefer the event's text and only fall back to the synchronous read when the
platform handed over none (chords, image-only clipboard, context-menu paste).
A regression test pins that a paste event carrying text never triggers the
synchronous read.
Submenus of the main menu were not constrained to the window (constrain: false, so they sit
beside their row), so the Object submenu ran past the bottom edge with its lowest commands out of
reach. Menus and submenus are now held within the window's height (moved up, never above the top,
8 pt margins, still beside their row so menu-aim is unchanged), and a panel taller than that
scrolls with the wheel; the keyboard's focus scrolls into view.
Model-Effort: Opus 5.5 drove; Opus 5.5 briefed; Sonnet 5.5 wrote at medium
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
* **No blank canvas** when switching back to a document tab or to a page in a file: the image it had when last shown is drawn from the first frame while it renders again
* Images of up to four recently shown files and pages are kept; closing a tab frees its image right away
Refs: #2074
pans_and_edits_get_no_preview_and_build_on_the_kept_image failed in full lib runs with "a pan after a previewed zoom renders strips only: Full". The render kept from the zoom is stale once the global font generation changes (retained.rs prepare: last.fonts == layout_generation()), and the once-per-process system font scan that many other tests start bumps it (fontdb.rs scan_system_fonts -> clear_layout_cache). When that scan landed between the test's zoom and its pan, the pan redrew in full instead of patching. Unrelated to the clock or the threshold. document() now runs scan_system_fonts() first (a concurrent caller waits for the scan), so the generation is settled before any render. Test-only change; the shipped staleness check is intended behaviour.
Stress, full lib binary with 16 test threads: 3 of 6 runs failed before, 0 of 25 after.
Model-Effort: Opus 5.5 drove; Opus 5.5 briefed; Sonnet 5.5 wrote at medium
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
* fix(collab-net): never send the server key over plain ws:// to another machine
Publishing a plugin, uploading or sharing with the server's key accepted a
ws:// address for a remote host, so the reusable administrative key went
out in clear text. The key now goes only over wss://, or over ws:// to
this machine (loopback or localhost); browsing without a key is unchanged.
* address review: connect only to loopback addresses for a plain ws://localhost endpoint
* fix(ui): access preferences from home screen via shortcut and sidebar (#956)
* fix(ui): handle preferences shortcut before popup guard
* docs: add docstrings for modified functions to satisfy coverage check
---------
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Tests for the -129 leftward boundary, the vertical boundaries both ways, a drag with no move along the flow, and several moving layers using their combined box. Doc comments say whose edge is "leading" and what the caller passes; the parity row names the unchanged cases and the web-app measurement source.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 triaged and briefed; Sonnet 5.5 verified and wrote at medium
A stale "field had the keyboard" flag no longer ends the tool on the first frame back from another tab. A selection appearing after K with nothing selected no longer moves the keyboard from the canvas to the Scale field. Ctrl+Enter and Alt+Enter are not the tool's exit. Tests pin the full text selection, the 2x and 1.5 factors, the field letting go after Enter and Escape, idle frames after Tab, and a leftover edit dropped on return.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 triaged and briefed; Sonnet 5.5 verified and wrote at medium
Choosing Crop in the colour picker starts crop mode only for one layer's
own image fill. The new test chooses Crop for a stroke, for two layers'
fills at once, and for a region's fill in vector edit mode (the layer's
own fill an image too). In each case it checks that the paint takes the
Crop mode, crop mode does not start, and the picker stays open. It also
checks that crop_on_canvas gives nothing for a solid fill, an index with
no fill, the page background and a style.
* Naming: record Developer Mode in the feature-names doc and the stage paragraph
docs/feature-names.md gains the entry for the rename: the old label, the new one, Figma's
US registration for DEV MODE (7574348) as the reason with its TSDR record, the identifiers
that keep their names, and the translations that moved from a generic "dev mode" to the
developer mode. The stage paragraph in ROADMAP.md listed the mode as Dev Mode.
* Actions palette: put each search doc comment back on its function
search_text had grown a second doc line under matches's, so the sentence about the query sat
above a function that takes none. search_text now says what it reads and matches keeps its own.
Adds ta.toml with every UI string, command label and menu wording, marked draft.
Desktop draws Tamil with an installed face (#2018); the web picker leaves Tamil out
until craft-fonts has Noto Sans Tamil, like Georgian and Arabic, with a test.
docs/translating.md gains Tamil's entry, typography notes and a Tamil draft section.
BFlat